One of the important tracks leading to natural risk prevention, disaster mitigation or the reduction of losses due to natural hazards is the vulnerability assessment of an "at-risk" region. The majority of researchers propose to assess vulnerability according to an expert evaluation of several qualitative characteristics, scoring each of them usually using three ratings: low, average, and high. Unlike these investigations, we attempted a quantitative vulnerability assessment using multidimensional statistical methods. Cluster analysis for all 89 Russian regions revealed five different types of region, which are characterized with a single (rarely two) prevailing factor causing increase of vulnerability. These factors are: the sensitivity of the technosphere to unfavorable influences; a "human factor"; a high volume of stored toxic waste that increases possibility of NDs with serious consequences; the low per capita GRP, which determine reduced prevention and protection costs; the heightened liability of regions to natural disasters that can be complicated due to unfavorable social processes. The proposed methods permitted us to find differences in prevailing risk factor (vulnerability factor) for the region types that helps to show in which direction risk management should focus on.